Modulation of Task-Evoked Pupil Dilations Differs in Individuals with Autism Spectrum Disorders: Implications for Noradrenergic and Cholinergic Dysfunction in ASD

Background: It has been posited that cognitive deficits among individuals with autism spectrum disorders (ASD) are a consequence of an imbalance in cortical excitation and inhibition. The noradrenergic and cholinergic systems have been implicated in globally modulating the homeostasis of excitatory and inhibitory neural activity and, accordingly, regulating cognitive processes such as attention, memory, and learning. Importantly, the noradrenergic system is critical for attending to task-relevant stimuli, a behavior that can be challenging for individuals with ASD. Whether there are inherent differences in the release of norepinephrine and/or acetylcholine in individuals with ASD, and the potential impact of these differences on their behaviors, remains unclear.

Objectives: The goal of this study was to determine whether, relative to matched typically developed participants, individuals with ASD exhibit differences in task-evoked pupil dilations under conditions of cognitive load. Any such difference might implicate atypicalities in the regulation of the noradrenergic and cholinergic systems in ASD.

Methods: 17 participants with ASD (15 males, 2 females; mean ± SD age = 32 ± 7.6 years) and 14 neurotypical controls (12 males, 2 females; mean ± SD age = 27 ± 4.5 years) performed a one-back working memory letter detection task (with stimuli presented every 0.5 s), while pupil size was measured at a rate of 1000 Hz. In half of trials, participants were also exposed to distractor auditory tones played at random intervals throughout the block. Feedback on hits and false alarms was displayed to participants in real time.

Results: A repeated-measures ANOVA revealed no significant group differences on either sensitivity indices or reaction times, indicating comparable performance between the two groups. A support vector machine was trained using cross-validation on the impulse response function (IRF) of pupil size in response to hits and was able to classify the two groups on diagnosis with above-chance accuracy. To evaluate differences in the magnitude of pupil size fluctuations between the two groups, a repeated-measures ANOVA was performed on the response amplitudes of the pupil size IRFs in response to hits, false alarms, and misses, for both task conditions (with/without distractor stimuli). There was a significant main effect of group on pupil response amplitudes, as well as a significant interaction of group by task condition. Post-hoc analyses indicate that whereas the pupil response amplitudes are not different across conditions for controls, for participants with ASD, the pupil response amplitudes were significantly lower when distractor stimuli were present compared to when they were absent.

Conclusions: These findings suggest that measurement of task-evoked pupil dilations allows for classification of individuals with ASD from neurotypical controls. Notably, when individuals with ASD successfully perform a cognitive task, the extent of pupil size fluctuation is nonetheless smaller when distractor stimuli are present. It is thus possible that individuals with ASD might exhibit atypical activity of the noradrenergic and cholinergic systems when attempting to attend to task-relevant stimuli.